1. Field of the Invention
The present invention relates to an image processing apparatus and an image processing method that generate output image data by carrying out image processing on input image data such as the processing of enhancing or suppressing the edge by adjusting the output levels of the pixel values in the outline region of characters or line drawings, etc., and/or halftone processing in order to reproduce a halftone.
2. Description of Related Art
At the time of printing out halftone characters or graphics in a printer, etc., it is common to carry out screen processing when reproduction of halftone is given priority. However, when halftone text characters are subjected to screen processing using a large screen, there was the problem that the apparent resolution gets reduced and the outline regions of the text characters become indistinct. Regarding this problem, a method has been disclosed (for example, see Japanese Unexamined Laid-Open Patent Publication No. Hei 9-282471) that judges the feature of each pixel (the type such as character, graphic, etc.) from the information described in the page description language (hereinafter referred to as PDL) and carries out image processing according to that feature. According to this method, for example, the text signal indicating that the pixel belongs to a text character is generated for pixels in the image area of text where priority is usually given to the resolution, and during image processing, a high resolution screen such as a line screen of 400 lines, for instance, is used for the image area for which this text signal has been set, while a low resolution screen such as a line screen of 200 lines, for instance, is used for other areas, thus avoiding the degradation in the resolution of the text part by using screens of different resolutions.
However, in the above method, a blank part is generated near the border where the screen is changed, and this appeared like an outline thereby causing degradation of the image quality. Further, while resolution is given importance than the tone reproduction characteristics in the case of text characters, etc., there are cases when the tone reproduction characteristics need to be given priority, such as when the character size is large, and in such cases, it is likely that the required tone reproduction characteristics cannot be obtained even when a high resolution screen such as a 400-line single line screen is used.
Compared to this, in Japanese Unexamined Laid-Open Patent Publications No. Hei 9-18710 and No. 2001-239706, for instance, a method has been proposed in which the type of image data specified by PDL is recognized and the image data is separated into the outline information indicating text characters, line drawings, and outline parts, and the halftone information indicating halftone areas, and different type of processing is made separately for these two types of areas, by carrying out high resolution bit map processing for the outline information and screen processing for the halftone information. Because of this, it becomes possible to reproduce halftone text characters with enhanced outline sections.
However, since in the above method it is necessary to carry out conversion of resolution two times for at least one of the outline information and the halftone information, there is the problem that the configuration of the apparatus becomes complex. In addition, when several types of PDL are present, since means for separating the outline information will be necessary for each of the PDL languages at the time of separating the outline information, the configuration of the apparatus naturally becomes complex.
Furthermore, depending on the output characteristics on the printer side, there are cases where the widths of the characters or thin lines is reproduced thicker than the correct line widths. In this case, it is possible to suppress the outline portion and make the line of text characters thinner by carrying out the processing of reducing the output level of the outline part of the text characters. For example, a method has been proposed in Japanese Unexamined Laid-Open Patent Publication No. Hei 7-334672 in which the curved line parts with step changes and the straight line parts without step changes are identified by pattern matching, and smoothing processing is made for the curved line parts having step changes while line width adjustment is made for the straight line parts without step changes.
Here, since the output level of the outline parts of the text characters will have to be reduced at the time of carrying out conversion to thin lines in the case of text characters that are expressed in full density over the entire area (the so called ‘solid characters’ or ‘solid letter’), when thin line parts are contained in the text characters, the densities of those fine line parts will be reduced excessively, thereby causing the problem of these lines to appear as if they are broken. In order to solve this problem, in the method described in the JP672 Patent Publication, since it is possible to specify the line width by pattern matching, it is possible to take the countermeasure of excluding the line parts with thin line widths from processing of conversion to thin lines, thereby making it possible to prevent the loss of thin line parts caused by conversion to thin lines. However, the method described in the above mentioned '672 patent Publication can only be applied to the case where the input image has binary values, and it is not possible to take the countermeasures of enhancing the outline or converting to thin lines for halftone text characters or line drawings.